Co-expression of three MEP pathway genes and geraniol 10-hydroxylase in internal phloem parenchyma of Catharanthus roseus implicates multicellular translocation of intermediates during the biosynthesis of monoterpene indole alkaloids and isoprenoid-derived primary metabolites


  • Vincent Burlat,

    1. EA 2106, Plant Biocompounds and Biotechnology, UFR des Sciences et Techniques, Université de Tours, 37200 Tours, France,
    2. Unité sous Contrat reconnue par l'INRA ‘Facteurs de transcription et ingénierie métabolique végétale’, Université de Tours, Tours, France
    Search for more papers by this author
  • Audrey Oudin,

    1. EA 2106, Plant Biocompounds and Biotechnology, UFR des Sciences Pharmaceutiques, Université de Tours, 37200 Tours, France, and
    Search for more papers by this author
  • Martine Courtois,

    1. EA 2106, Plant Biocompounds and Biotechnology, UFR des Sciences Pharmaceutiques, Université de Tours, 37200 Tours, France, and
    Search for more papers by this author
  • Marc Rideau,

    1. EA 2106, Plant Biocompounds and Biotechnology, UFR des Sciences Pharmaceutiques, Université de Tours, 37200 Tours, France, and
    Search for more papers by this author
  • Benoit St-Pierre

    Corresponding author
    1. EA 2106, Plant Biocompounds and Biotechnology, UFR des Sciences et Techniques, Université de Tours, 37200 Tours, France,
    2. Unité sous Contrat reconnue par l'INRA ‘Facteurs de transcription et ingénierie métabolique végétale’, Université de Tours, Tours, France
      For correspondence (fax +33 247 36 70 42; e-mail
    Search for more papers by this author

For correspondence (fax +33 247 36 70 42; e-mail


In higher plants, isopentenyl diphosphate (IPP) is synthesised both from the plastidic 2-C-methyl-d-erythritol 4-phosphate (MEP) and from the cytosolic mevalonate (MVA) pathways. Primary metabolites, such as phytol group of chlorophylls, carotenoids and the plant hormones abscisic acid (ABA) and gibberellins (GAs) are derived directly from the MEP pathway. Many secondary metabolites, such as monoterpene indole alkaloids (MIAs) in Catharanthus roseus, are also synthesised from this source of IPP. Using Northern blot and in situ hybridisation experiments, we show that three MEP pathway genes (1-deoxy-d-xylulose 5-phosphate synthase (DXS), 1-deoxy-d-xylulose 5-phosphate reductoisomerase (DXR) and 2C-methyl-d-erythritol 2,4-cyclodiphosphate synthase (MECS)) and the gene encoding geraniol 10-hydroxylase (G10H), a cytochrome P450 monooxygenase involved in the first committed step in the formation of iridoid monoterpenoids display identical cell-specific expression patterns. The co-localisation of these four transcripts to internal phloem parenchyma of young aerial organs of C. roseus adds a new level of complexity to the multicellular nature of MIA biosynthesis. We predict the translocation of pathway intermediates from the internal phloem parenchyma to the epidermis and, ultimately, to laticifers and idioblasts during MIA biosynthesis. Similarly, the translocation of intermediates from the phloem parenchyma is probably also required during the biosynthesis of hormones and photosynthetic primary metabolites derived from the MEP pathway.


In higher plants, the common building blocks for the synthesis of several thousand isoprenoid-derived primary and secondary metabolites is isopentenyl diphosphate (IPP) and its isomer dimethylallyl diphosphate (DMAPP). The two respective pathways leading to the biosynthesis of IPP are the cytosolic mevalonate (MVA) and the plastidic 2-C-methyl-d-erythritol 4-phosphate (MEP) pathways (Lichtenthaler, 1999; Lichtenthaler et al., 1997; Rodríguez-Concepción and Boronat, 2002; Figure 1). Among the primary metabolites derived from the MEP pathway are the plant hormones gibberellins (GAs) and abscisic acid (ABA), and photosynthesis actors such as chlorophylls (phytol group) and carotenoids (Kasahara et al., 2002; Lichtenthaler, 1999; Lichtenthaler et al., 1997; Figure 1). Secondary metabolites, such as the terpenoid moiety of monoterpene indole alkaloids (MIAs) in Madagascar periwinkle (Catharanthus roseus), are also derived from the MEP pathway (Chahed et al., 2000; Contin et al., 1998; Veau et al., 2000; Figures 1 and 2). MIAs form a large group of structurally diverse metabolites, some of which possess pharmaceutical value. For example, vinblastine and vincristine are potent antitumoral agents, whereas ajmalicine is used in treatment against hypertension (Levêque et al., 1996). In C. roseus aerial organs, MIAs appear to accumulate in specialised cells, the laticifer–idioblast system (St-Pierre et al., 1999). Recent studies showed that the last two steps in the biosynthesis of vindoline, one moiety of the active dimers vinblastine and vincristine, occur in these MIA-accumulating cells (St-Pierre et al., 1999; Figure 2). Interestingly, the central steps of this pathway, involving the formation of tryptamine and secologanin, and their subsequent condensation to form strictosidine, occur in the epidermis (Irmler et al., 2000; St-Pierre et al., 1999; Figure 2). These results, obtained through in situ hybridisation and immunolocalisation experiments, showed that certain intermediates must be transported from the epidermis to the laticifer–idioblast cells (St-Pierre et al., 1999). Here, we investigated the cell-specific expression of four genes implicated in upstream steps of the long MIA biosynthetic pathway in C. roseus: three MEP pathway genes, 1-deoxy-d-xylulose 5-phosphate synthase (DXS), 1-deoxy-d-xylulose 5-phosphate reductoisomerase (DXR) and 2C-methyl-d-erythritol 2,4-cyclodiphosphate synthase (MECS), as well as the gene encoding geraniol 10-hydroxylase (G10H), the first monoterpenoid-dedicated enzyme.

Figure 1.

Biosynthetic pathways for IPP showing their subcellular organisation and main isoprenoid products in higher plants.

DXS, 1-deoxy-d-xylulose-5-phosphate (DXP) synthase; DXR, DXP reductoisomerase; MECS, 2C-methyl-d-erythritol-2,4-cyclodiphosphate (MEC) synthase; G10H, geraniol 10-hydroxylase; HMGR, 3-hydroxy-3-methyl-glutaryl-CoA reductase.

Figure 2.

Biosynthetic pathway for monoterpenoid indole alkaloids in C. roseus aerial organs showing the multicellular compartmentation of gene expression.

Among the nine genes already studied, three MEP pathway genes (DXS, DXR and MECS), as well as G10H, are expressed in internal phloem parenchyma (this study). Three intermediate genes are then expressed in epidermis (SLS, secologanin synthase (Irmler et al., 2000); TDC, tryptophan decarboxylase and STR, strictosidine synthase (St-Pierre et al., 1999)), whereas late genes are further expressed in specialised laticifer (black arrowhead) and idioblast (white arrowhead) cells (D4H, desacetoxyvindoline 4-hydroxylase and DAT, deacetylvindoline 4-O-acetyltransferase; St-Pierre et al., 1999).

In this paper, we report a similar spatial and developmental expression pattern for these four genes in the aerial organs of C. roseus. Interestingly, these genes are co-expressed in internal phloem parenchyma cells, a tissue not previously implicated in MIA biosynthesis. This surprising result is discussed with respect to both MIA and isoprenoid primary metabolite biosynthesis. In both cases, we suggest that the translocation of metabolites from phloem parenchyma towards different tissues is necessary to accomplish the biosynthesis of the various isoprenoid-derived primary and secondary metabolites.


Organ-specific gene expression

The expression pattern of DXS, DXR, MECS and G10H in several C. roseus organs was studied by RNA blot hybridisations (Figure 3). RNA transcripts of these four genes were present in all tissues analysed, except in flowers and fruits, where G10H transcripts were hardly detectable (Figure 3). Overall, the four genes showed a coordinated expression profile. The accumulation of mRNA was consistently higher in roots and in young and actively growing aerial organs (flower buds, young leaves, first internodes) as compared to mature organs (flowers, mature leaves, fruits). This correlates with the organ-specific localisation of MIA biosynthesis described previously by St-Pierre et al. (1999).

Figure 3.

Northern blot analysis of the DXS, DXR, MECS and G10H transcripts in C. roseus organs.

Lower panel shows rRNA ethidium bromide staining. R, roots; FB, flower buds; F, mature flowers; Fr, fruits; IN, first internodes; YL, young developing leaves taken on the first node; ML, mature leaves taken on the second node; OL, older leaves taken on the third node. Note the overall coordinated expression profile of the four genes. Level of transcripts is higher in roots and in young and actively growing aerial organs (flower buds, young leaves, first internodes) compared to mature organs (flowers, mature leaves, fruits).

In situ hybridisation

To study the cell-specific expression pattern of these four genes, in situ hybridisation experiments were performed on several aerial tissues of C. roseus. Labelling with the DXS antisense probe in the vasculature regions of longitudinal sections of young developing leaves (Figure 4a) was clearly located above the xylem in the adaxial phloem region under dark-field observation (Figure 4b). Under these conditions, the xylem displayed a typical white birefringence, whereas the label appeared pink-red. Thus, in situ hybridisation experiments performed on serial sections of C. roseus organs allowed accurate comparison of the expression pattern for DXS, DXR, MECS and G10H. The same pattern of expression was obtained using the four antisense probes (Figure 4c–f), with no significant background observed with the corresponding sense probes as negative controls (Figure 4g–j). Interestingly, shorter development times allowed two regions in the adaxial phloem to be distinguished: one with labelled cells and the other with unlabelled cells, the latter being located between the first region and the tracheary elements (Figure 4k). The DXS-expressing cells were clearly parenchymatous. First, they were broadly round in shape (i.e. compare cross- and longitudinal sections; Figure 4k,l) and they possessed chloroplast-like structures (Figure 4k); second, they presented the typical red fluorescence of chlorophylls when comparing whole-mount petiole sections with the DXS antisense labelling in a similar cross-section (Figure 4m–o). Indeed, the pink-red signal revealing DXS gene expression (Figure 4m) co-localised to the adaxial part of the vascular bundle, associated with the red autofluorescence of chlorophyll (Figure 4n). In contrast, the unlabelled conducting adaxial and abaxial phloem regions had a lower level of chlorophyll autofluorescence (Figure 4n,o). Consistent with the Northern blot results, the four genes presented a coordinated pattern of expression, both developmentally and spatially (Figure 3). This was confirmed by in situ hybridisation experiments performed on sections of various organs (Figure 5). The presented results show the localisation of DXS expression, but similar results were obtained for DXR, MECS and G10H antisense probes hybridised to serial sections (data not shown). In a cross-section of the basal part of a young leaf, the hybridisation signal specifically localised to the adaxial phloem (Figure 5a) in correlation with the labelling observed on longitudinal sections (Figure 4). This specific labelling was seen at very early stages of development, even in the cotyledons of young seedlings (Figure 5b). Similarly, the hybridisation signal was detected in the internal phloem parenchyma of the stem located in the outer pith border of the youngest internodes just under the shoot apical meristem (Figure 5c,d). In flower buds, labelling was observed in the internal phloem parenchyma located in the floral tube and sepals (data not shown), as well as within carpels (Figure 5e) and petals (Figure 5f,g). In general, higher levels of signal were observed in the youngest parts of various organs. Interestingly, in young developing leaves, the signal followed a basipetal gradient, decreasing from the young, revoluted base towards the older leaf tip (Figure 6a). A similar gradient of expression was observed with other MIA-specific genes in epidermis (Figure 6b) and laticifer idioblasts (Figure 6c).

Figure 4.

Cell-specific expression of three MEP pathway genes (DXS, DXR and MECS) and G10H in C. roseus young developing leaves and petioles.

(a–j) Serial longitudinal sections of the revoluted leaf base.

(k) Cross-section of the central vein in a developing leaf.

(l) Longitudinal section of a secondary vein in a developing leaf.

(m) Paraffin-embedded cross-section of the petiole of a developing leaf.

(n, o) Hand-cut fresh cross-section of a similar petiole as in (m).

Paraffin-embedded sections of young developing leaves were hybridised with digoxigenin-labelled transcripts. Hybridised transcripts were localised with antidigoxigenin–alkaline phosphatase conjugates followed by BCIP–nitroblue tetrazolium colour development. Hybridisation with antisense RNA probes: (a–c, k–m) DXS; (d) DXR; (e) MECS; (f) G10H. Control hybridisation with sense RNA probes: (g) DXS; (h) DXR; (i) MECS; (j) G10H. The photographs of the hybridisation are in bright field except for (b,m) which are in dark field, and (l) which is in bright field with phase contrast. (n) autofluorescence observed under a blue-violet excitation filter set; (o) bright field. Note the perfect co-expression of the three MEP pathway genes and G10H with various levels of relative intensity (c–f) and the absence of non-specific labelling with the corresponding negative controls (g–j). (b) Dark field observation of (c) allows to distinguish the red birefringent hybridisation signal in the adaxial (internal) phloem area above the white birefringent signal of the tracheary elements of the xylem. (k–l) Shorter colour development of two perpendicular orientations allow to clearly localise the hybridisation signal in adaxial (internal) round-shaped phloem parenchyma cells above the conducting elements of the adaxial (internal) phloem. The presence of plastids in the MEP pathway specific cells, distinguishable in (k) is confirmed by the presence of chlorophyll occurring in these cells as seen by their red autofluorescence (n), and their greenish aspect in bright field (o). ipp, adaxial internal phloem parenchyma; ipc, adaxial internal phloem conducting elements; te, tracheary elements of xylem; epc, abaxial external phloem conducting elements; white arrows show ipc and white arrowheads show epc. Bars = 100 µm.

Figure 5.

Cell-specific expression of DXS in internal phloem parenchyma of C. roseus aerial organs.

(a) Cross-section of the base of a young leaf.

(b) Longitudinal section of cotyledons from 5-day-old seedling.

(c) Longitudinal section of the apical meristem and first internode regions (note that the grey signal observed in the apical meristem is not specific, it is also observed with the sense probe in a serial section; data not shown).

(d) Cross-section of the first internode region of the stem.

(e) Longitudinal section of a carpel.

(f) Longitudinal section of a flower bud in the petal region.

(g) Cross-section of petals in a flower bud.

Sections were hybridised as described in Figure 4 legend. Note that the results presented show the localisation of DXS expression, but similar results were obtained for DXR, MECS and G10H antisense probes hybridised to serial sections. Bars = 200 µm except for (b), which is 100 µm.

Figure 6.

Comparison of the basipetal gradient of expression of MIA-specific genes in three different cell types of C. roseus leaves.

(a) Expression of DXS gene in internal phloem parenchyma.

(b) SLS gene expression in epidermis.

(c) DAT gene expression in laticifer–idioblast cells.

Serial longitudinal sections of a 8-mm leaf were hybridised as described in Figure 4 legend. Despite the fact that different tissues are labelled, the overall intensity of labelling shows a similar gradient from the young revoluted base towards the older leaf tip. Note that a similar gradient was observed with the alkaloid-specific genes tested in Figure 2. Bars = 1 mm.


MEP pathway is expressed in internal phloem parenchyma cells

We have shown that the biosynthesis of monoterpenoid precursors in aerial parts of C. roseus involves a cell type of the internal phloem. Three MEP pathway biosynthetic genes are expressed in a specific group of parenchyma cells associated with internal phloem tissue, as shown by the accumulation of the gene transcripts in this cell type throughout the aerial parts (Figures 4 and 5). Like all Apocynaceae species, C. roseus develops an additional internal phloem during its histogenesis, which is also called intraxylary phloem in stems and adaxial phloem in leaves. However, no specific function for the associated parenchyma cells has been defined. In stems, these parenchyma cells are not morphologically differentiated from ground parenchyma of the pith, but instead they are just located in the periphery of the pith next to the intraxylary conducting elements of the phloem. In leaves, the DXS-, DXR- and MECS-expressing cells, located above the adaxial sieve tube area, form a delimited group with a characteristic bean shape. In both organs, the presence of chloroplast in internal phloem parenchyma cells is fully consistent with the plastidic location of the MEP pathway enzymes (Araki et al., 2000; Bouvier et al., 2000; Carretero-Paulet et al., 2002; Lichtenthaler, 1999; Lichtenthaler et al., 1997). Although DXS, DXR and MECS represent only three of the seven biosynthetic enzymes in the MEP pathway, we suggest that IPP production is restricted to these cells. This notion is supported by the co-localisation of G10H transcripts, the first committed step in the formation of the monoterpenoid moiety of MIAs (Collu et al., 2001), which are known to be derived from the MEP pathway. We coin the name ‘IPP cells’ (internal phloem parenchyma cells) to evoke the function of this group of distinct cells in IPP biosynthesis.

Although Arabidopsis thaliana does not develop an internal phloem, MEP pathway gene expression has been localised to vascular tissue in this species. The expression of DXS and DXR was mostly detected along the veins of young developing leaves and cotyledons as shown by studies of promoter-GUS constructs in transgenic plants (Carretero-Paulet et al., 2002). Expression was also present in mesophyll tissues but to a lower extent. In contrast, DXS, DXR and MECS transcripts are not found in the mesophyll cells of C. roseus. Although, we cannot rule out the possibility that in situ hybridisation is not sensitive enough to detect low level of transcripts in this tissue, it is also possible that this represents a genuine difference between both species.

It is notable that several plant species have two class of DXS gene, including one (class 2) that was suggested to be preferentially involved in secondary metabolism (Walter et al., 2002). The preferential expression of C. roseus DXS, a class 2 gene based on sequence comparison in leaf and roots (Figure 1), is consistent with a role of these DXS genes in terpenoid-accumulating organs/cells in several species. However, the next six steps in the MEP pathway are encoded by single genes in Arabidopsis (Lange and Ghassemian, 2003) and in the rice genome (personal observation). Therefore, localisation of DXS, together with DXR and MECS transcripts, is likely to show any of the cells that actively express MEP pathway genes.

Some isoprenoid-derived primary metabolism involves different cell types

Several primary metabolites are derived from the MEP pathway. For example, plastid isoprenoids involved in photosynthesis, such as carotenoids and the phytol moiety of chlorophylls, are synthesised from the MEP pathway (Lichtenthaler et al., 1997). At first sight, the specific expression of MEP pathway genes in C. roseus internal phloem tissues is rather surprising as photosynthetic pigments are mostly localised in mesophyll tissues. One cannot exclude the fact that high level of mRNA is not a good reflection of the level of enzyme activity in a particular cell type, or that mRNA or protein itself is transported to another site of activity. However, our data, based on the localisation of four different genes involved in the MEP pathway and terpenoid metabolism (G10H), suggest that internal phloem parenchyma is the main site of isoprenoid biosynthesis and implies the intercellular translocation of metabolites.

Recent results suggest that phytol biosynthesis in leaf might involve translocation of an isoprenoid precursor from the vascular tissue in Arabidopsis. The progression of chlorophyll accumulation from the vein to mesophyll tissue in cla (chloroplastos alterados)1 mutant, supplemented or not with mevalonate, suggests that a vascular-tissue-derived precursor is translocated to the mesophyll cells (Nagata et al., 2002).

The plant hormones ABA, which is derived from carotenoids, and GAs (Kasahara et al., 2002) are also products of the MEP pathway. Particularly surprising is the localisation of these hormone biosynthetic pathways in such a way that individual genes are expressed in different cell types. An intermediate step in ABA formation, short-chain dehydrogenase/reductase, is highly restricted to specialised vascular tissues that are distinct from the tissue-specific expression pattern of other ABA pathway genes in Arabidopsis, indicating that intercellular translocation may be required for the completion of ABA biosynthesis (Cheng et al., 2002). In addition, studies of GA biosynthesis in Arabidopsis and tobacco also predicted such translocation processes. Genes involved in early steps of GA biosynthesis, such as ent-kaurene synthase A and copalyl diphosphate synthase, show expression in (pro)vascular tissues compatible with the MEP pathway localisation, whereas GA 3β-hydroxylase and GA 3-oxydases, encoding enzymes catalysing final steps, are expressed in non-vascular tissues at the target sites (Itoh et al., 1999; Silverstone et al., 1997; Yamaguchi et al., 2001).

The implication of multiple cell types and the intercellular translocation of pathway intermediate may be a more common notion in isoprenoid metabolism than previously appreciated.

Four cell types are involved in MIAs biosynthesis

It is notable that the developmental- and organ-specific expressions of MEP pathway genes and G10H are coordinately regulated with latter steps in MIA pathway in C. roseus. The level of transcripts of DXS, DXR, MECS and G10H is higher in young and actively growing tissues than in mature organs (Figure 3). Within a single leaf, a basipetal gradient of expression was observed (Figure 6). In addition, gene expression was also high in roots where MIAs accumulate. This is consistent with the results reported by St-Pierre et al. (1999) and Irmler et al. (2000), showing the same organ-specific distribution and expression gradients for several downstream genes in MIA biosynthesis.

However, the cell-specific regulation of MEP pathway genes and G10H is distinct. The formation of tryptamine and secologanin, and their subsequent condensation to form strictosidine occur in epidermal cells of young developing shoots and leaves of C. roseus, whereas the latter steps leading to the formation of vindoline are localised within specialised alkaloid-accumulating cells, i.e. the laticifers and idioblasts (Figure 2; Irmler et al., 2000; St-Pierre et al., 1999). These results predict the translocation of an intermediate metabolite from the epidermis to the laticifer–idioblast system. In this report, we show that the expression of three genes involved in the MEP pathway and G10H are co-localised in an unexpected cell type – the internal phloem parenchyma. This indicates that a specific cell type is involved in the initial stages of monoterpenoid biosynthesis, and implies a second translocation of an intermediate metabolite between 10-hydroxygeraniol and loganin (Figure 2). Participation of multiple cell types in the MIA pathway and the intercellular translocation of intermediates emerges as a common design in alkaloid biosynthetic systems (Bird et al., 2003; De Luca and St-Pierre, 2000).

Internal phloem function in the biosynthesis and/or accumulation of various secondary metabolites

In addition to monoterpenoid biosynthesis, the internal phloem in C. roseus and other Apocynaceae is the site of biosynthesis or accumulation of other secondary metabolites. Flavonoid 3',5' hydroxylase has been immunolocalised in internal phloem of young C. roseus leaves and flower buds (Kaltenbach et al., 1999). In contrast, chalcone synthase, an early enzyme in the flavonoid pathway, was detected in the upper epidermal layer. Interestingly, these results also suggest an intercellular transport occurring between epidermis and internal phloem during flavonoid biosynthesis, although in the opposite direction as compared to MIAs pathway.

Toxic phytosterols accumulate in the internal phloem of Apocynaceae species according to the following circumstantial evidence. Members of this family are known to contain cardiac glycosides, which are toxic phytosterols synthesised by the plant as an effective defence strategy (Fraenkel, 1959). The aphid Aphis nerii is known to collect cardiac glycoside in order to produce bright yellow pigment and toxic defence compounds, possibly protecting both the insect and the plant against predators. Interestingly, A. nerii preferentially feeds from internal adaxial phloem of Apocynaceae species, suggesting that cardiac glycosides are localised in this tissue (Botha et al., 1975, 1977).

Building complex secondary metabolites, such as the monoterpenoid indole alkaloids, requires the biosynthesis of several precursors, and the assembly and chemical decoration of intermediates. Participation of multiple cell types with specialised metabolic function and the intercellular movement of intermediates emerge as a common design of these biosynthetic factories. Internal phloem appears to be a highly specialised site playing a pivotal role in formation of precursors and decoration of monoterpenoids. This implies that several mechanisms of transport must exist to deliver isoprenoid intermediates out of the phloem parenchyma plastids and to other compartments and/or cells where they will be further metabolised or stocked. A more general role for this cell type could emerge as further isoprenoid pathways are mapped at the cellular level.

Experimental procedures

Northern blots

Tissues samples were taken from 2-month-old C. roseus that were immediately frozen in liquid nitrogen. Total RNA was extracted using an RNA isolation kit (Plant RNeasy extraction kit, Qiagen, Courtaboeut, France). For Northern blot analysis, total RNA (15 µg) was fractionated by electrophoresis in a 1.5% (w/v) agarose gel containing 2.2 m formaldehyde and blotted onto Hybond N+ membrane (Amersham-Biosciences, Orsay, France). For probe preparation, the full-length cDNAs DXS, DXR, MECS and G10H, excised from pGEM-T vector, were α-32P-labelled with the ‘Prime-a-gene’ labelling kit (Promega, Charbonnières, France). Hybridisation was carried out for 18 h at 42°C in Ultrahyb hybridisation solution (Ambion, Cambridgeshire, UK). Washes were performed for 10 min at room temperature once in 0.1× saline sodium phosphate EDTA (SSPE) and 0.5% (w/v) SDS and for 30 min at 42°C two times in 0.1× SSPE and 0.5% (w/v) SDS.

Tissue fixation, embedding and sectioning

RNase-free conditions were strictly observed for all steps. All glasswares were baked for 8 h at 180°C. Mature C. roseus, grown in greenhouse, were harvested in late spring–early summer, and rapidly fixed in formaldehyde acetic acid alcohol (FAA) and embedded in Paraplast as previously described by St-Pierre et al. (1999). Serial sections (10 µm) were spread on silane-coated slides overnight at 40°C, and paraffin was removed using xylene (two times for 15 min) before re-hydration in an ethanol gradient series up to diethylpyrocarbonate (DEPC)-treated water.

In situ RNA hybridisation

Full-length cDNA cloned in pGEM-T (Promega) was used for the synthesis of sense and antisense RNA probes. DXS cDNA clone contained a 2387-bp fragment (Chahed et al., 2000; GenBank Accession number AJ011840). DXR cDNA clone contained a 1740-bp fragment (Veau et al., 2000; GenBank Accession number AF250235). MECS cDNA clone contained a 994-bp fragment (Veau et al., 2000; GenBank Accession number AF250236). G10H (CYP76B6) cDNA clone contained a 1560-bp fragment (Collu et al., 2001; GenBank Accession number AJ251269). For DAT and SLS, previously described plasmids were used (Irmler et al., 2000; St-Pierre et al., 1999). Plasmids were linearised with 5′ overhang restriction enzymes (Promega) and used as a template for RNA probe synthesis with digoxigenin-UTP and SP6 or T7 RNA polymerases (Promega), according to the manufacturer's instruction (Roche, Molecular Biochemicals, Meylon, France). Probes were hydrolysed at 60°C for 13–19 min, depending on the transcript size, in order to obtain around 400-bp fragments (Jackson, 1992).

Hybridisation protocol was essentially as in St-Pierre et al. (1999) with a few modifications. Unless otherwise stated, all steps were performed at room temperature. Before hybridisation, re-hydrated sections were treated with proteinase K (5 µg ml−1 in 100 mm Tris–HCl and 50 mm EDTA (pH 8.0)) for 30 min at 37°C, followed by two rinses with TBS150 (10 mm Tris–HCl, 150 mm NaCl (pH 7.5)), by blocking of proteinase K with glycine (2 mg ml−1 in TBS150) for 2 min, and by two rinses in TBS150. Sections were post-fixed with 3.7% formaldehyde in PBS for 20 min and washed two times in TBS150 for 5 min each. Finally, sections were acetylated with acetic anhydride (0.25% in 0.1 m triethanolamine–HCl (pH 8.0)) for 10 min, washed with TBS150, dehydrated in an ethanol series and air-dried. For hybridisation, aliquot of hybridisation mixture (120 µl) was dispersed on the sections and mounted under cover slips (22 mm × 50 mm) to prevent evaporation. Hybridisation mixture included 200 ng ml−1 partially hydrolysed digoxigenin-labelled RNA transcripts, 40% formamide, 10% dextran sulfate, 1 µg ml−1 de-proteinised yeast total RNA, 0.3 m NaCl, 0.01 m Tris–HCl (pH 6.8), 0.01 m sodium phosphate (pH 6.8), 5 mm EDTA and 40 units ml−1 RNasin ribonuclease inhibitor (Promega). Hybridisation was for 16–20 h at 50°C in an atmosphere of 50% formamide. Cover slips were then removed by soaking in 2× SSC at 37°C (1× SSC is 0.15 m NaCl and 0.015 m sodium citrate). Slides were treated with RNase A (50 µg ml−1 in 10 mm Tris–HCl, 0.5 m NaCl, 1 mm EDTA (pH 7.5)) for 30 min at 37°C, and then washed under gentle agitation in 2× SSC for 1 h, in 1× SSC for 1 h and in 0.1× SSC for 1 h at 65°C. For immunolocalisation of hybridised transcripts, slides were washed in Tween TBS (TTBS) (0.1 m Tris–HCl (pH 8.0), 0.15 m NaCl and 0.3% Triton X-100) for 10 min and blocked with 2% BSA fraction V (Sigma Aldrich Chemie, Lyon, France) in TTBS for 30 min. Sheep antidigoxigenin Fab fragments–alkaline phosphatase conjugate (Roche) at a 1 : 200 dilution in a solution of 1% BSA in TTBS was dispensed on the sections and mounted under cover slips. After incubation for 2 h, the unbound conjugates were washed two times for 15 min with TTBS and two times for 10 min with AP buffer (0.1 m Tris–HCl (pH 9.5), 0.1 m NaCl and 10 mm MgCl2). For colour development, slides were immerged in 175 µg ml−1 5-bromo-4-chloro-3-indolyl phosphate (BCIP) and 350 µg ml−1 nitro blue tetrazolium chloride in AP buffer for 1–16 h. After development, slides were washed in water, dried and mounted in immersion oil under cover slips.


Sections were observed with an epifluorescence microscope (Olympus BX51) equipped with a digital camera (Olympus DP50) and the corresponding software (olympus analysis). Most images were captured under bright field. In some cases, in order to highlight xylem tracheary elements, either bright field with phase contrast or dark field observations were performed. Chlorophyll autofluorescence was imaged with a blue-violet excitation filter set (400–440 nm excitation filter, 475 nm cutting filter). The wide-view low-magnification pictures were reconstructed from several images using the multiple image alignment function of the analysis software.


This research was financially supported by the Ministère de l'Education Nationale, de la Recherche et de la Technologie (MENRT, France) and by the Ligue contre le Cancer comité d'Indre et Loire). We thank Dr J. Memelink (University of Leiden, the Netherlands), who kindly provided C. roseus G10H cDNA.